A new study seems to provide a better way to categorize El Nino climate events, and offers an explanation for how different kinds of El Nino events emerge.
El Nino is part of a large scale, very important climate phenomenon in the Pacific Ocean, generally referred to as the El Nino Southern Oscillation (ENSO). Over time (years) wind and water currents move heat into the upper levels of the Equatorial Pacific (La Nina). Then, over time (months) the heat comes back out – that is an El Nino. The effects can be dramatic. During El Nino years, trade winds and monsoons may behave differently than normal. How much precipitation falls and where it falls can change over large regions. Deserts become lakes, good croplands are drought stricken, sea levels change across large portions of the coast.
It is interesting to contemplate the following thought experiment (sorry, a bit of a digression). Imagine if all of the conditions associated with El Nino happened all the time, and had been happening for centuries. An El Nino that is always there is not really an El Nino. It is normal. Those parts that are dry would be dry; Plants and animals, including people, would be dry adapted there in physiology, ecology, and behavior. Same for wet places. It wouldn’t be a desert covered with a lake, it would just be a lake. It wouldn’t be a drought, but just a desert. Etc. The point of this is to underscore the real meaning of El Nino: change. It isn’t so much what it does, but rather, that this climate event’s effects are sudden, dramatic, and occasional.
El Nino has been off and on in the news over the last year or so because it looked like there was going to be a really big one in 2014, but it never materialized. (Even without an El Nino, which warms the surface of the Earth, 2014 was still a record breaking warm year.) Now, El Nino is in the news again because finally we kinda sorta are having one, and a future (this year and next) super double El Nino is being predicted.
Why is El Nino prediction so difficult, and why, when an El Nino happens, it may be very different from some other El Nino that happened before in its overall intensity and in the details of what it causes to happen elsewhere in the world?
You can hear them screaming. The climatologists. “Why? Why? WHY?!?!?” Because this is a really a big thing and it would be really nice to be better at predicting it.
A new study has taken an important step in understanding, and ultimately, predicting El Nino. “Strong influence of westerly wind bursts on El Nino diversity” by Chen et al, published in Nature Geoscience, makes two related points. First, the authors presupposed the existence of three kinds of El Ninos. It has long been thought that El Ninos can be classified into different categories, but the number and nature of those categories varies across groups of researchers. I asked the author if they tried using other a priori numbers for the El Nino categories. “Yes, we did try using other cluster numbers,” Dr. Chen told me. “If it’s set to 2, we would have the extreme El Nino and a broad cluster that include both the canonical and the warm-pool El Nino. If it’s set to 4, we would still see the 3 types we identified but with a 4th type that’s not well separated from the canonical El Nino. In any case we had only one type of La Nina. These discussions will be included in a long paper to be submitted to Journal of Climate.”
Chen et al used a method of modeling El Nino that is different than what is usually done and with this method successfully classified all of the El Nino events over a 50 year time period into these three categories. Second, Chen et al show that the main variable that determines what kind of El Nino happens is the intensity and location of westerly wind bursts (WWBs). I also asked if other variables used in their model (discussed below) were changed to see what would happen. Dr. Chen noted, “we did play with different model settings and parameters, and the outcome turned out to be fairly robust. We are very confident with our results.”
First a brief note on the method. The usual way of managing the complex phenomenon of El Nino … of measuring stuff and stuffing the measurements into a mathematical model … is called empirical orthogonal function (EOF) analysis. This involves measuring key variables across a grid covering the Pacific Equatorial region. Then you take the measurements and simplify how they are organized and turn a multidimensional time-space problem in to a one with fewer dimensions. There are different ways to do this but they all fall into the widely used methodology that included principle component analysis and other things you may be familiar with. You take something really complicated and derive simplified (somewhat) data that is more usable for characterizing a phenomenon or predicting the phenomenon’s behavior while at the same time not throwing away too much of the meaningful variation in the system. This method, however, if fairly linear and deterministic. A bunch of variables are thought to cause outcome A (which has variants), and this bunch of variables are combined so you have only X and Y causing A.
Chen et al applied a different (but well established) technique that presumes less about the linear nature of the model’s components and allows for complex interrelationships that may vary across conditions to remain. It is called fuzzy clustering method. In this method, the data are allowed to decide on their own (more or less) how they should be organized, and (this is the fuzzy part) individual bits of data are actually allowed to occupy more than one cluster. For many systems, the two methods would result in similar outcomes, but when a system is less linear the second method may be more realistic.
When this method is used, the role of WWBs turns out to be very important. This is not entirely new because we already knew that westerly winds across the Equatorial Pacific were important in ENSO cycles. The ENSO cycle involves, to simplify a bit, heat at the surface of the Pacific moving westward and then into the deep (but not to deep) ocean where it builds up. This process is maintained by currents and winds moving from east to west. It is a little like the air near your ceiling growing ever hotter if you burn wood in your stove; In that case the property of warm air rising causes the upper few feet of your living room to get much hotter than the floor. The Western Pacific gets hotter over time because winds and currents push the heat there.
This build up in heat (and other factors) eventually cause a change in the movement of heat and we see warm water moving east, surfacing, and transferring heat energy into the air. That is an El Nino.
From the abstract of the paper:
We propose a unified perspective on El Nin?o diversity as well as its causes, and support our view with a fuzzy clustering analysis and model experiments. Specifically, the interannual variability of sea surface temperatures in the tropical Pacific Ocean can generally be classified into three warm patterns and one cold pattern, which together constitute a canonical cycle of El Nin?o/ La Nin?a and its different flavours. Although the genesis of the canonical cycle can be readily explained by classic theories, we suggest that the asymmetry, irregularity and extremes of El Nin?o result from westerly wind bursts, a type of state-dependent atmospheric perturbation in the equatorial Pacific. Westerly wind bursts strongly affect El Nin?o but not La Nin?a because of their unidirectional nature. We conclude that properly accounting for the interplay between the canonical cycle and westerly wind bursts may improve El Nin?o prediction.
The authors demonstrate that accounting for WWBs does a better job of retroactively predicting the different kinds of El Nino events that have happened over the last fifty years. They conclude that El Nino may result from a combination of the built in see-saw effect of build up of ocean heat in the west and the reversal of movement of warm water on one hand and WWB perturbations, with the pattern of westerly winds affected by the oscillation itself. (I am over simplifying ENSO here, see below for resources on how it works.) Whether or not an El Nino happens is predicted by the classic oscillation model, but which kind of El Nino results is better predicted by the WWBs. From the study:
Such a scenario is appealing because it reconciles hotly debated issues related to the classification and genesis of various El Nin?o events, by killing three birds — diversity, asymmetry and extremes — with one stone. But one must not dwell on the simplicity of the picture painted here. Our intention is to emphasize the strong influence of WWBs on El Nin?o diversity, but not to downplay other processes that may play significant roles in El Nin?o dynamics and thus contribute to the complexity of its diversity.
The research reported here does not address, but may relate to, a set of questions that have been on my mind as I’ve watched El Nino and the discussion surrounding it develop over the last year or so. Is El Nino (or ENSO, more broadly) changing because of climate change? Since El Nino was already hard to predict, we can chalk up this last round of lousy predictions as El Nino being El Nino. But we might also ask the question, is it possible that as more surface and upper ocean heat enters the system, are there changes? Chen et all actually do note that “… real-time El Nin?o forecasting remains an elusive and formidable goal. This is probably because predictability estimates were mainly based on models dominated by a single mode of El Nin?o variability or on hindcast skills of relatively large El Nin?o events, whereas in reality El Nin?o has a variety of flavours, especially in the past decade” (emphasis added). So, these folks, referring to other research, note that El Nino has changed. Is this random variation with no important linear time dimension, or is it a “new normal” for the already normal-defying El Nino, or, perhaps, is it the first part of a period during which ENSO changes dramatically to a climate controlling phenomenon that acts differently in important ways?
Michael Tobis, an expert on atmospheric and ocean systems, suggested to me that “…the real action in climate change is where the warm water goes, not what the wind does. The wind will respond, and may reinforce or mitigate what the ocean does, true. But as the ocean water mass gets further from its recent near-equilibrium, eventually wind stress coupling becomes a smaller deal and the water will go where it will go.” Tobis also notes, and Chen et al acknowledge this may be important, that “the most salient feature in the oceans right now is the large and persistent warm blob in the eastern North Pacific.” This implies (i.e., causes me to speculate or, really, guess) that a warming ocean may shift the balance of what is important in driving, or resulting from, ENSO dynamics. That does not detract from Chen et al’s apparent ability to both classify and explain the differences between El Nino events with WWBs being the key factor.
I asked Dr. Chen to go out on a limb a bit to discuss what the future may hold as climate changes.
Question: With global warming, ocean heat (both at some depth and SST) has increased. Since heat in the ocean is a key variable in ENSO cycles, is it possible that El Nino dynamics would change in some important way, for example, of the three flavors of El Nino, the relative likelihood of which flavor manifesting changing? Is there evidence that such a change may have already occurred, thus the dismal level of predicability of 2014/5? Or, would you expect such a change in the future? My gut feeling is that El Nino dynamics is a barely stable metastable system that is in sufficiently weak equilibrium that it could change to a different equilibrium if important inputs are changed a lot.
Answer: I think your gut feeling is right, in the sense that El Nino is changing under global warming. The questions are how and why. Observations over the last 15 years seem to indicate that the system is now dominated by the warm-pool El Nino, while some people use IPCC model projections to argue that in the future the extreme El Nino will become more frequent. These are still open questions.
Second version of the same question: I’ve heard El Nino/ENSO described as a quasi equilibrium. The essential feature of this system is shifting back and forth between recharge and discharge of ocean heat. Is a different system imaginable where this is not a cyclic system, but rather, a steady state system (such as we see with the Atlantic Conveyor or other climate systems) with heat going in (somewhere) and coming out (somewhere else) more or less steadily? Since we are entering global temperature levels not seen in a long time (and thus only represented in ancient paleo records of lower quality) it seems like it can’t be ruled out (other than it being a rather extreme idea)
Answer: In the recent history and perhaps also during many periods in the past, ENSO did behave like a self-sustaining oscillation. However, it is quite possible that the system might enter a steady state — a permenant El NIno or La Nina state — when external forcing changes.
Question: Figure 4 (see top of post) seems to show something rather astonishing (aside from that figure’s use to demonstrate WWB and WWV association with different kinds of El Nino): WWB in 2014 was very high and uniquely so. Why? Other than the apparent fact that this WWB was not followed by a strong El Nino (a key point of your paper) is there anything else interesting about this?
It all depends on the interplay between the WWB and the basic cycle (measured by WWV). Only when the former occurs at the right phase of the latter a large El Nino will take place. It is true that WWBs were very strong in 2014, but only in the early year, not over the entire spring season. Further experiments will be needed to clarify whether or not the relationship between WWB and WWV will change under global warming.
It will certainly be interesting to see, over the next 24 months or so, if we end up having a strong El Nino, a double El Nino, or if we have lapsed into an extended period of what some are calling El Annoyingo.
For more information about El Nino:
Check out: The First Earth Day, an epoch journey into politics, explosions, folk music, and old boats floating on stinking rivers.